Sunday, September 20, 2009

MORE dimensionality

Often the solution to a prediction problem (model the supposed underlying phenomena using variables and relationships) requires a model with more dimensionality.

We have created a flatland where people are constantly trying to reduce dimensionality to that which is easily comprehensible, thereby missing the opportunity to develop models which SOLVE THE PROBLEM.

I posit that merely observing the data you have collected regarding some phenomena, using more dimensional visualization (while still comprehensible)... can only help in the above process.

This is what Glasshouse is about.

Arkowitz